Predictors of Job Satisfaction among New MSWs: The Role of Organizational Factors

Abstract

This study examines job satisfaction at early career stages among new U.S. Master of Social Work (MSW) graduates. It explores (a) what factors were associated with job satisfaction, including organizational factors (interpersonal working environment and agency characteristics), educational preparation in the MSW program, and personal characteristics, and (b) whether organizational factors have more significant effects than other predictors. Eighty graduates of a northeastern MSW program were surveyed. Organizational factors, specifically interpersonal working environment (atmosphere and quality of working relationship, satisfaction with general supervision) and agency characteristics (worker’s perceived effectiveness of service delivery in the agency) were correlated with job satisfaction. Educational preparation, defined as graduates’ perceived level of attainment of competence, was correlated with job satisfaction. Regression analysis revealed that two organizational factors (atmosphere and quality of working relationship and effectiveness of service delivery) were more significant predictors than other characteristics. To help neophyte social workers better transition into the reality of the workplace, agency efforts to build positive working relationships with supervisors and coworkers are recommended. Workers’ confidence in the effectiveness of services should be enhanced through an emphasis on evidence-based practice. Further studies that develop measurements of early-stage worker competence and examine more diverse factors are recommended.

Keywords:

• Social work education
• job satisfaction
• early career in social work
• new social work graduates

Notes

1 We tested whether multiple regression was appropriate by examining the data and the residuals from the equation. The sample size (80) was slightly below the 84 recommended for analyses with 4 independent variables and medium effect sizes (Tabachnick & Fidell, 2001; Soper, 2017). We accepted the risk associated with low sample size because recommended rules of thumb vary dramatically by discipline (between 10 and 200 in this example) (Cross Validated, 2016). The analyses for multicollinearity indicated no problems in the relationships among independent variables. The independent variables were moderately correlated to each other (between .26 and .61); the tolerances or proportion of variance not explained by other independent variables were .59 or higher in all instances (.1 or lower is considered problematic); and the highest variance inflation factor (VIF) was 1.89 where values above 5.0 may be problematic.

For analyses of the residuals (errors) from the final regression equation (Table 2), the standardized residuals were normally distributed, suggesting good model fit. Scatterplots of predicted values on residuals were linear but not rectangular, indicating some problems with homoscedasticity (constant variance).